D O C U M E N T 5 3 7 M AY 1 9 2 9 4 6 1 537. To Chaim Herman Müntz [Berlin,] 27 May 1929 Dear Mr. Müntz, I thank you for sending the calculations. It will be important to know how many equations are equivalent to the condition [1] that is still unclear to me. Regarding the problem as a whole, Lanczos’s discovery changes the situation fundamentally.[2] My opinion is now oriented around the following considerations: It is impossible that the equations hold in general, since these equations alone already satisfy causality. Therefore, cannot be the case. I begin with the field equations and decompose them into From (I), it follows that from this and from the identity , by subtraction we obtain the equation … (III). It is quite natural to consider (I) as the “gravitational equations,” (III) as the elec- tromagnetic field equations. The parenthesis in (III) is then the electric field strength, . Then, however, we cannot set the antisymmetric equations (II), since they would require the vanishing of the field (pure gravitational field). (III) is—this is the im- portant point—derived from (I) alone. These are, however, only 10 independent equations. We would still certainly have to add the equations , if that is permitted, i.e., if there is no overdetermina- tion. I am not sure about that. Then, the terms with would vanish, while would be a temporarily unknown constant. The gravitational equations in vacuum would then differ from the only in terms of second order, which should be permis- sible. It would in any case be essential that . With best greetings, your A. Einstein S v  0 = R i 0 =  1  2 0 = = G  =  1 G *  2 G ** = + + 0 (I) (symmetric) =  1 G *+ 2G** 0 (II) (antisymmetric) = D   1 G *  2 G ** +   0 = D   G *+ 1 2 G*  0  D  1G* 2G** +  0 = f  S   0 =  1  2 R ik 0 =  2 0 